firebird-mirror/src/msgs/build_file.epp

/*
 *	PROGRAM:	JRD Message Facility
 *	MODULE:		build_file.epp
 *	DESCRIPTION:	Build message file
 *
 * The contents of this file are subject to the Interbase Public
 * License Version 1.0 (the "License"); you may not use this file
 * except in compliance with the License. You may obtain a copy
 * of the License at http://www.Inprise.com/IPL.html
 *
 * Software distributed under the License is distributed on an
 * "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, either express
 * or implied. See the License for the specific language governing
 * rights and limitations under the License.
 *
 * The Original Code was created by Inprise Corporation
 * and its predecessors. Portions created by Inprise Corporation are
 * Copyright (C) Inprise Corporation.
 *
 * All Rights Reserved.
 * Contributor(s): ______________________________________.
 *
 * 2002.02.15 Sean Leyne - Code Cleanup, removed obsolete "XENIX" port
 * 2002.02.15 Sean Leyne - Code Cleanup, removed obsolete "DELTA" port
 *
 * 2002.10.30 Sean Leyne - Removed support for obsolete "PC_PLATFORM" define
 *
 */

#include "firebird.h"
#include <stdio.h>
#include "../jrd/common.h"
#include <stdlib.h>
#include <string.h>
#include "../jrd/ibase.h"
#include "../jrd/common.h"
#include "../jrd/msg.h"
#include "../jrd/gds_proto.h"

#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif

const int max_levels = 4;
typedef msgnod* msgnod_ptr_array[max_levels];

DATABASE DB = "msg.fdb";

static void ascii_str_to_upper(TEXT*);
static USHORT do_msgs(const TEXT*, const TEXT*, bool);
static void propagate(msgnod**, msgnod**, ULONG, ULONG);
static SLONG write_bucket(const msgnod*, USHORT);
static void sanitize(TEXT*);

static SLONG global_file_position;
static int global_file;

/* keep the LOCALE_PAT names in sync with file_params.h */
// CVC: It's #define MSG_FILE_LANG	"intl/%.8s.msg"
// in jrd/file_params.h, it seems we are already out of sync!

#ifdef WIN_NT
#include <io.h> // open, lseek, write, close
const char* const locale_pattern = "%.8s.msg";
const char path_separator = '\\';
#else
const char* const locale_pattern = "%.10s.msg";
const char path_separator = '/';
#endif

const char* const FILENAME	= "firebird.msg";

#ifdef VMS
#include <file.h>
#else
#include <sys/types.h>
#ifndef WIN_NT
#include <sys/file.h>
#endif
#endif  /* VMS */

#ifndef O_RDWR
#include <fcntl.h>
#endif

#ifndef O_BINARY
#define O_BINARY	0
#endif


int CLIB_ROUTINE main( int argc, char** argv)
{
/**************************************
 *
 *	m a i n
 *
 **************************************
 *
 * Functional description
 *	Top level routine.
 *
 **************************************/
	TEXT db_file[MAXPATHLEN], filename[MAXPATHLEN], pathbuffer[MAXPATHLEN];
	BASED ON LOCALES.LOCALE this_locale;

	strcpy(db_file, "msg.fdb");
	strcpy(filename, FILENAME);
	TEXT* pathname = NULL;
	TEXT* locale = NULL;

	bool sw_warning = false;
	const TEXT* const* const end_args = argv + argc;

	for (++argv; argv < end_args;) {
		const TEXT* p = *argv++;
		bool sw_bad = false;
		if (*p != '-')
			sw_bad = true;
		else
		{
			switch (UPPER(p[1])) {
			case 'D':
				strncpy(db_file, *argv++, MAXPATHLEN);
				db_file[MAXPATHLEN - 1] = 0;
				break;

			case 'F':
				strncpy(filename, *argv++, MAXPATHLEN);
				filename[MAXPATHLEN - 1] = 0;
				break;

			case 'P':
				strncpy(pathbuffer, *argv++, MAXPATHLEN);
				pathbuffer[MAXPATHLEN - 1] = 0;
				pathname = pathbuffer;
				break;

			case 'L':
				locale = *argv++;
				break;

			case 'W':
				sw_warning = true;
				break;

			default:
				sw_bad = true;
			}
		}
		if (sw_bad) {
			printf("Invalid option \"%s\".  Valid options are:\n", p);
			printf("\t-D\tDatabase name\n");
			printf("\t-F\tMessage file name\n");
			printf("\t-P\tMessage file path (should end with '/')\n");
			printf("\t-L\tLocale name\n");
			printf("\t-W\tVerbose warning messages\n");
			exit(FINI_ERROR);
		}
	}

/* Get db_file ready to go */

	READY db_file AS DB;
	START_TRANSACTION;

/* make sure the path name ends in a '/' */

	if (pathname) {
		size_t len = strlen(pathname);
		//fb_assert(len + 1 < MAXPATHLEN);
		if (pathname[len - 1] != path_separator) {
			pathname[len] = path_separator;
			pathname[len + 1] = 0;
		}
	}

/* check for the locale option  */

	if (!locale) {				// no locale given: do the regular US msgs
		if (!pathname)
			do_msgs(filename, NULL, false);
		else {
		    //fb_assert(strlen(pathname) + strlen(filename) < MAXPATHLEN);
			strcat(pathname, filename);
			do_msgs(pathname, NULL, false);
		}
	}
	else {
		bool got_one = false;
		FOR FIRST 1 L IN LOCALES WITH L.LOCALE = locale
			got_one = true;
		END_FOR;
		if (got_one) {			// do only 1 locale
			strcpy(this_locale, locale);
			sanitize(locale);
			if (pathname) {
				sprintf(filename, locale_pattern, locale);
				strcat(pathname, filename);
				do_msgs(pathname, this_locale, sw_warning);
			}
			else
				do_msgs(filename, this_locale, sw_warning);
		}
		else {
			strncpy(this_locale, locale, sizeof(this_locale));
			ascii_str_to_upper(this_locale);
			if (!strcmp(this_locale, "ALL")) {
				FOR LC IN LOCALES // do all locales except piglatin
					WITH LC.LOCALE != "pg_PG" AND LC.LOCALE != "piglatin";
					locale = LC.LOCALE;
					strcpy(this_locale, LC.LOCALE);
					printf("build_file: building locale %s", this_locale);
					sanitize(locale);
					sprintf(filename, locale_pattern, locale);
					if (pathname) {
						strcat(pathname, filename);
						printf(" to file %s\n", pathname);
						do_msgs(pathname, this_locale, sw_warning);
					}
					else {
						printf(" to file %s\n", filename);
						do_msgs(filename, this_locale, sw_warning);
					}
					pathname = NULL;
				END_FOR;
			}
			else {
				printf("build_file: Unknown locale: %s\n", locale);
				printf("Valid options are:\n");
				FOR LO IN LOCALES
					printf("\t%s\n", LO.LOCALE);
				END_FOR;
				printf("\tall\n");
			}
			COMMIT;
			FINISH DB;
			exit(FINI_ERROR);
		}
	}

	COMMIT;
	FINISH;

	return(FINI_OK);
}


static void ascii_str_to_upper( TEXT* s)
{
/**************************************
 *
 *	a s c i i _ s t r _ t o _ u p p e r
 *
 **************************************
 *
 * Functional description
 *	change the a - z chars in string to upper case
 *
 **************************************/

	while (*s) {
		*s >= 'a' && *s <= 'z' ? *s &= 0xDF : *s;
		*s++;
	}
}


static USHORT do_msgs( const TEXT* filename, const TEXT* locale, bool sw_warning)
{
/**************************************
 *
 *	d o _ m s g s
 *
 **************************************
 *
 * Functional description
 *	Build the message file
 *
 **************************************/

	// Divy up memory among various buffers
	// Memory leaking until the program finishes?
	msgrec* leaf_node = (MSGREC) gds__alloc((SLONG) MSG_BUCKET);
	msgrec* const leaf = leaf_node;
	const TEXT* const end_leaf = (TEXT*) leaf + MSG_BUCKET;

	// Open output file

	global_file = open(filename, O_WRONLY | O_CREAT | O_BINARY, 0666);
	if (global_file == -1) {
		printf("Can't open %s\n", filename);
		return FINI_ERROR;
	}
	global_file_position = 0;

	// Format and write header
	ISC_MSGHDR header;
	header.msghdr_major_version = MSG_MAJOR_VERSION;
	header.msghdr_minor_version = MSG_MINOR_VERSION;
	header.msghdr_bucket_size = MSG_BUCKET;
	// CVC: Since this is an unused field that holds garbage in the *.msg file,
	// we'll initialize it to the date the FB project was registered with SF.
	header.msghdr_origin = 20000730; // 2000-07-30
	write_bucket((msgnod*) &header, sizeof(header));

	// Write out messages building B-tree
	
	int warning_counter = 0;
	USHORT n = 0;
	ULONG position = 0, prior_code = 0;
	TEXT msg_text[256];
	msgnod_ptr_array buckets, nodes;
	nodes[0] = NULL;
	size_t len = 0;

	FOR X IN MESSAGES SORTED BY X.CODE

		// pre-load with US English message - just in case we don't
		// have a translation available.

		strcpy(msg_text, X.TEXT);
		len = strlen(msg_text);
		if (locale) {			// want translated output
			// Overwrite English message with translation, if we find one
			int found_one = 0;
			FOR Y IN TRANSMSGS WITH Y.FAC_CODE = X.FAC_CODE
				AND Y.NUMBER = X.NUMBER
					AND Y.LOCALE = locale
					AND Y.TEXT NOT MISSING AND Y.TEXT != " ";
				strcpy(msg_text, Y.TEXT);
				len = strlen(msg_text);
				found_one++;
			END_FOR;
			if (!found_one && sw_warning)
				printf
					("build_file: Warning - no %s translation of msg# %ld\n",
					 locale, X.CODE);
			if (found_one > 1 && sw_warning)
				printf
					("build_file: Warning - multiple %s translations of msg# %ld\n",
					 locale, X.CODE);
			if (found_one != 1)
				warning_counter++;
		}

		if (leaf_node->msgrec_text + len >= end_leaf) {
			position = write_bucket((msgnod*) leaf, n);
			propagate(buckets, nodes, prior_code, position);
			leaf_node = leaf;
		}
		leaf_node->msgrec_code = prior_code =
			MSG_NUMBER(X.FAC_CODE, X.NUMBER);

		leaf_node->msgrec_length = len;
		leaf_node->msgrec_flags = X.FLAGS;
		n = OFFSETA(MSGREC, msgrec_text) + len; // useless? See assignment below.
		TEXT* p = leaf_node->msgrec_text;
		if (len) {
			const TEXT* q = msg_text;
			do {
				*p++ = *q++;
			} while (--len);
		}
		n = p - (SCHAR *) leaf;
		leaf_node = NEXT_LEAF(leaf_node);
	END_FOR;

	// Write a high water mark on leaf node

	if (leaf_node->msgrec_text + len >= end_leaf) {
		n = (SCHAR *) leaf_node - (SCHAR *) leaf;
		position = write_bucket((msgnod*) leaf, n);
		propagate(buckets, nodes, prior_code, position);
		leaf_node = leaf;
	}

	leaf_node->msgrec_code = -1UL;
	leaf_node->msgrec_length = 0;
	leaf_node->msgrec_flags = 0;
	n = (SCHAR *) leaf_node - (SCHAR *) leaf;
	position = write_bucket((msgnod*) leaf, n);

	// Finish off upper levels of tree

	header.msghdr_levels = 1;

	for (msgnod** ptr = nodes, **ptr2 = buckets; *ptr; ptr++, ptr2++) {
		msgnod* node = *ptr;
		node->msgnod_code = -1UL;
		node->msgnod_seek = position;
		n = (SCHAR *) (node + 1) - (SCHAR *) * ptr2;
		position = write_bucket(*ptr2, n);
		++header.msghdr_levels;
	}

	header.msghdr_top_tree = position;

/* Re-write header record and finish */

	lseek(global_file, LSEEK_OFFSET_CAST 0, 0);
	write(global_file, &header, sizeof(header));
	close(global_file);
	global_file = -1;

	if (warning_counter)
		printf("build_file: %d messages lack translations in locale %s\n",
				  warning_counter, locale);

	return FINI_OK;
}


static void propagate(
					  msgnod** buckets,
					  msgnod** nodes, ULONG prior_code, ULONG position)
{
/**************************************
 *
 *	p r o p a g a t e
 *
 **************************************
 *
 * Functional description
 *	Propagate a full bucket upward.
 *
 **************************************/
/* Make sure current level has been allocated */

	if (!*nodes) {
		*nodes = *buckets = (msgnod*) gds__alloc((SLONG) MSG_BUCKET);
		nodes[1] = NULL;
	}

/* Insert into current bucket */

	msgnod* node = (*nodes)++;
	node->msgnod_code = prior_code;
	node->msgnod_seek = position;

/* Check for full bucket.  If not, we're done */

	const msgnod* const end = (msgnod*) ((SCHAR*) *buckets + MSG_BUCKET);

	if (*nodes < end)
		return;

/* Bucket is full -- write it out, propagate the split, and re-initialize */

	position = write_bucket(*buckets, MSG_BUCKET);
	propagate(buckets + 1, nodes + 1, prior_code, position);
	*nodes = *buckets;
}


static SLONG write_bucket( const msgnod* bucket, USHORT length)
{
/**************************************
 *
 *	w r i t e _ b u c k e t
 *
 **************************************
 *
 * Functional description
 *	Write stuff, return position of stuff written.
 *
 **************************************/
	const USHORT padded_length = ROUNDUP(length, sizeof(SLONG));
	const SLONG position = global_file_position;
	int n = write(global_file, bucket, length);
	if (n == -1) {
		fprintf(stderr, "IO error on write()\n");
		exit(FINI_ERROR);
	}
	
	SLONG zero_bytes = 0;
	n = write(global_file, &zero_bytes, padded_length - length);
	if (n == -1) {
		fprintf(stderr, "IO error on write()\n");
		exit(FINI_ERROR);
	}
	global_file_position += padded_length;

	return position;
}


static void sanitize( TEXT* locale)
{
/**************************************
 *
 *	s a n i t i z e
 *
 **************************************
 *
 * Functional description
 *	Clean up a locale to make it acceptable for use in file names
 *	  for Windows NT, PC and mpexl: remove any '.' or '_' for mpexl,
 *	replace any period with '_' for NT or PC.
 *	  Keep this in sync with gds.cpp
 *
 **************************************/
	while (*locale) {
		const SSHORT ch = *locale;
		if (ch == '.')
			*locale = '_';
		locale++;
	}
}